PSYFR

Pax Trax

Main difficulties in Airports today are the inexorable rise in passengers and resulting congestion. The oxymoron, is that it`s becoming more difficult to simplify the passenger journey, optimise resources, and maximise NAR in sync.

Airport Terminals strive to become resource optimised processing boxes, but to make that work and improve pax UX, requires ever increasing intelligence and data insight to support sophisticated operational decisions.

Pax Trax provide the answers you didn’t know existed to your most valuable commercial and operational questions. We capture and analyse “difficult to get” information, because getting more of the same data isn’t going to solve it.

Real Time Queue Management

The Challenge

Simple binary reporting of average queue times and poor performance SLA breach’s, is not what good looks like. It’s changed. It’s had to. Now it’s about trying to lessen the number of variable outcomes; good, is managing a consistent and efficient optimised security processing box, matching staff resource to demand slot peaks improving pax UX and NAR and hopefully, in sync.

The Process

Our data collection toolkit [Pax Trax] captures location data, but also matched to their flight destination(s), available airside dwell time and impact upon NAR; and provide a near 100% reliable SLA breach alert.

The KPI dashboard refresh rate is 180 seconds and not based upon data sampling to provide an “average” queue time, because, averages lie.

Operationally relevant our KPI dashboard features an Experiment Suite that aks, could that “operational decision” been made differently? And for that read better and what would have been the outcome? Unique data function also include RCT Random Control Tests affords managers the opportunity of comparing the performance of two completely flight processing performance side by side measuring all key data input variables.

The Results

Pax Trax accurately measures queue times for passengers, and operations by “process point” aligned to your SLA. Combined with sophisticated predictive analytical techniques for capacity forecasting we have improved several major UK Airports departure security processing functions through:

Reduced SLA Breach`s

Reduced “average processing cost per pax per hour” by 5%.

Increased pax flow rate through security channel(s) by 6% through optimisation and pax behavioural segmentation.

Capacity Forecasting, Flow Management
and Resource Optimisation

The Challenge

The disconnect between what’s forecast and what really happens is because historical data is a poor indicator of future events; compounded by weak statistical analysis and “shallow data sets” are insufficient to make confident decisions.

Confidence in capacity and passenger flow forecasts start strong, but confidence evaporates when it plays out in the terminal. With fixed assets and staff rotas the only “flex” in the terminal is a lengthening queue time. With a baked in the cake xxx pax per hour processing rate per channel you rely on a constant uniform pax flow even in demand slot peaks.

The most variable factor is the queue time. Every decision feeds into that. The resource optimisation needle will only “flicker on the dial” if your pax forecasts can be relied upon. There is a direct performance correlation between the representativeness of predicted flow rate and your ability to plan resources to process those passengers. That`s the unlock.

The process

Forecasting needs to be uncoupled from historical trend analysis as the basis of resource allocation and operational decisions. Machine learning, predictive analytics and adopting time series analysis to produce daily forecasts with additional contextual data inputs will provide management with a heightened confidence. It`s about decision support and confidence in the numbers day after day.

The results

A 3.5% saving in operating costs in demand slot peaks, based upon average processing cost per pax per hour in a UK Top 10 Airport.

Experiment Suite & Data Science

An Experiment Suite takes the guesswork out of large scale implementations by allowing you to plan structured experiments throughout the Airport to test hypotheses and theories in a scientific way. The experiment suite allows you to run a series of scenarios to establish whether a better ROI could have been delivered.

Data Science

Key to achieving maximising value from the data generated every day in your airport are the data science and artificial intelligence algorithms used in modern business. Our Data Science Hub provides a suite of tools integrating popular algorithms and packages into a single environment. Rather than transferring vast quantities of data between databases and software stacks, the Data Science Hub provides everything required for analysis.

RCT’s. Random Control Tests

Even good decisions can be plagued by angst over the alternative not taken.

Randomised controlled trials to test different interventions and the impact they have on customer behaviour. A “deep dive analysis” of previously unmeasurable data comparing two (2) completely different flights from the same airport on the same day, but differing airlines (a scheduled business flight viz a viz a holiday flight). By inputting different input variables i.e. 100% of pax afforded a 10% increase in dwell time- what was the effect upon spend?

Asynchronous data capture

We have uniquely perfected an asynchronous data capture system which sync`s and merges data (automatically) between say a passenger’s smart device and i.e. an ABCR gate or epos till. This process unlocks value from previously un measurable data insights.

We can significantly exceed and infill your gaps in purchasing behaviour and spending by passengers once airside. This cascade of previously unobtainable data unlocks opportunities to improve UX and NAR. No longer rely on historical WDF epos data, average dwell time and guess work.

By x tabbing data we construct socio demographic cluster groups that exhibit certain characteristics allowing you to understand how these pax, consume, interact and behave by flight, TOD, destination etc. This could for example allow you to adopt variable pricing models if using digital screen in F&B and incentivise spend?

Male Passenger In Airport Departure Lounge Scanning Digital Boarding Pass On Smart Phone

F+B spend analysis by flight

The missing jigsaw piece in measuring NAR is attributing F&B spend to individual flights. Unlike WDF sales, airports cannot apportion spend to individual pax as there is no way of knowing end destination, there is no “identifier”. Our system uses the common denominator (smart device) to which we attach some of the flight details at the ABCR gates.

Tokenisation ensures privacy. With an additional epos black box connected to the restaurant till we can sync purchases to mobile knowing its flight destination. The spend data captured provides the same details as printed on till receipt. We provide the ground truth, itemised spend in F&B, by flight not average spend or guess work.

Life Time Value (LTV) Modelling

LTV Modelling measures the “value” to the airport of individual passenger group clusters. The “value” is the complete sum of all the important data inputs that affect the passenger`s journey and goes way beyond the typically narrow sub set available to airports.

The LTV model measures and ranks entirely new data sets that were previously unobtainable challenging deeply held anecdotal beliefs or observational data that somehow are indelibly ingrained into airports understanding.

The model has many inputs including loyalty (to the airport), not based upon spurious “satisfaction” mkt research (unreliable), but considering the true catchment area of your current passengers.

Smart Trax for Security

Cyber Security

Network traffic control
and Software defined network

We address two classes of cyber-security problem.

Firstly, autonomous devices that form part of the Internet of things may never have their software updated, leaving them increasingly vulnerable to attack as new exploits for them are found. They commonly have a long and mostly unmonitored service life to further increase that risk and are already widely used unnoticed by their owners in DDoS attacks. Such devices typically have few or no threat management features built in and are very unlikely to have such protection retrofitted.

Secondly, the traditional methods of securing computer equipment of all kinds are evidently not sufficient. This is partly because they respond to threats rather than pre-empt them, so new exploits will always have some time to work. Our objective is to add a new method of threat management that is light on human effort, heavy on automation, and attempts to mitigate even unknown threats. It is independent of existing cyber-security solutions, complementing rather than replacing them.

Network security is a complex multifaceted problem, created by an endless series of new and evolving threats.

Our network security solution handles some of the more difficult and recently emerging of those problem facets, such as protection for the Internet of things.

Core features use a sophisticated technique that is highly automated and independent of network architecture.

It can address whole classes of threat while still being able to confront very specific threats.

Optional components extend analysis and control to provide an even more sophisticated system.

Of course, to do something that sophisticated requires something complex.

The elegance of our solution is in the way we combine leading edge technologies with an objective focused interface this cleanly separates requirements from their implementation, making it relatively simple and natural to use.

Think of it as defining what you want to do, without any concern for the underlying mechanism.

Your intentions are codified in a hidden cascade of sophisticated behaviour that realises your objectives and informs you of factors that might adjust your thinking.

Ultimately, our network security solution makes your network more secure with little effort.

Human Presence Detection HPD

Brexit, Points Based Immigration System, COVID – 19, Global Warming, it’s only going one way. Human presence detection is the new technology revealing those not wanting to be seen. Never has the cost economics of prevention been higher or the numbers of people and traffickers trying to reach these shores illegally.

To protect borders from illegal immigrants, people smuggling and human trafficking we have developed a disrupter for the unseen. Concealment in lorries and metal shipping containers can out stretch current detection technique capabilities; it also requires the vehicle to be stationary.

We can scan moving vehicles for human presence and indicating the approximate location(s) of illegal immigrants in seconds. This is a multi tiered system blending HPD and Mobile phone detection as a secondary indicator of presence. HPD relies upon the distortion of low energy electromagnetic radiation and detecting a human “signature” presence which is unique even when concealment techniques such as silver foil, deep hiding between pallets of goods, chilled temperature and metal sided vehicles are deployed.

Information gathering
and security

Intelligent location analytics – remote sensor capability.

Our sensors capture data by sniffing the air for data packets released by smart devices. Not reliant on an app or Wi-Fi, and undetectable, we can scan both 2.4 + 5.0 GHz wave band simultaneously capturing virtually all data traffic. Sensor can be left or transported by drone to a remote location and left outside in all weathers {IP65]– capturing data for a week in a 150 mtr radius.

Using an anomaly algorithm selecting key data fields we can establish locational behavioural patterns “out of sync” with the general public as a bench mark. These sensors can placed out of sight and gather in close to real time intelligence at demonstrations. Making sense of the locational behaviour of people’s digital wash will reveal quicker than observation by CCTV of their movement and interaction with other people. Even if found information cannot be extracted from the sensor due to its secure configuration.